CN111045068B

CN111045068B - Low-orbit satellite autonomous orbit and attitude determination method based on non-navigation satellite signals

Info

Publication number: CN111045068B
Application number: CN201911383023.4A
Authority: CN
Inventors: 孙红星; 锁应博; 陈锐志; 肖雄武; 郭丙轩
Original assignee: Wuhan University WHU
Current assignee: Wuhan University WHU
Priority date: 2019-12-27
Filing date: 2019-12-27
Publication date: 2021-01-15
Anticipated expiration: 2039-12-27
Also published as: CN111045068A; AU2020103576A4

Abstract

The invention relates to a low orbit satellite autonomous orbit and attitude determination method based on non-navigation satellite signals, which comprises the following steps: the multi-sensor integration on-track calibration; during the operation of the satellite, adopting a corresponding orbit and attitude determination method according to the number of the control points and the matching points and whether a navigation satellite signal is available; as the satellite flies around, a global geographic feature point library is constructed by the existing geographic feature point library and the continuously updated geographic feature point library introduced by the SLAM, and the original feature point library is continuously updated in the running process of the satellite and the precision of the original feature point library is improved; and forming a global geographic feature point library, and finally realizing the orbit and attitude determination based on the non-navigation satellite signal INS/SS/SMNS. The invention avoids the infeasibility of global ground control point arrangement, and is simultaneously suitable for normal operation of the orbit and attitude determination system under the condition of abnormal navigation satellite in an extraordinary period. Particularly, the establishment of a global geographic feature point library can realize the high-precision orbit and attitude determination of INS/SS/SMNS, and greatly improve the safety of a spatial information network in an emergency.

Description

Low-orbit satellite autonomous orbit and attitude determination method based on non-navigation satellite signals

Technical Field

The invention relates to the technical field of satellite navigation, in particular to a low-orbit satellite autonomous orbit and attitude determination method based on non-navigation satellite signals.

Background

At present, there are 4 main technical means for determining the orbit and attitude of a low-orbit Satellite with high precision, including a Satellite Laser Ranging (SLR), a Doppler terrestrial radio positioning (dorsi), a precision Ranging and velocity measuring (PRARE) and a Global positioning (GNSS). However, SLR is expensive, heavy, and difficult to perform alone for the precise positioning task of low orbit satellites with a height of about 500km due to the limited observation coverage area and the severe influence of weather. The DORIS system in France has relatively slow data acquisition and processing speed, and the ground tracking network coverage is relatively weak. The PRARE system in Germany has fewer stations around the world, expensive equipment and a small number of satellites to load the system. However, GNSS cannot meet the requirements of orbit determination and attitude determination in extreme periods such as failure, and cannot realize full-arc orbit determination and attitude determination in working period, and there are many difficulties in the autonomous orbit determination and attitude determination method and system for low-orbit satellites based on non-navigation satellite signals.

With the continuous improvement of the construction of the Beidou satellite system, the spatial information network requires the capabilities of high-resolution earth observation, more accurate navigation and positioning and the like. For future spatial information network plans in China, most network nodes are located in low orbits and near-earth spaces, and for accurately unifying the network nodes to the space-time reference of the BDS in real time, a BDS/GNSS receiver needs to be carried, and accurate positions and speed parameters of the network nodes are determined through absolute and relative positioning modes. If the spatial information network node needs to realize centimeter-level real-time orbit accuracy index, firstly, the real-time orbit and clock error accuracy of the navigation satellite must be improved, and the current BDS and GPS cannot meet the condition. Therefore, it is very important to provide an autonomous orbit and attitude determination technology for a low orbit satellite based on a non-navigation satellite signal, which is not only suitable for special needs in an emergency, but also can provide redundant observation in a normal period.

The commonly used on-satellite orbit and attitude determination sensors at present comprise on-satellite GNSS receivers, inertial navigation, star sensors and the like, and with the continuous development of computer vision technology, new technologies such as SLAM and the like begin to rise, which widens the thought for on-satellite orbit and attitude determination technology. The SLAM technology can effectively solve the problem that ground control points required by orbit and attitude determination are difficult to arrange globally, and can be combined with other sensors to provide a high-precision orbit and attitude determination result. INS is well established and has gained a great deal of engineering practice. SS or INS/SS combination is widely used in satellite attitude determination, but the depth of the combination is not enough, and the accelerometer information cannot be applied. SMNS has also been widely used in the fields of aerial photogrammetry, missile guidance, and the like, and SLAM has been used in the fields of robot navigation, indoor navigation, and the like and has been rapidly developed. In this regard, these technologies have already had a good foundation in their original fields, and for autonomous orbit and attitude determination of low-orbit satellites, these technologies are combined by means of data fusion to achieve the effects of deeper depth, more reliability and higher precision. Therefore, the low-orbit satellite autonomous orbit and attitude determination method based on the non-navigation satellite signals (INS/SS/SLAM/SMNS combined technology) is significant for the condition that the navigation signals are lacked and control points are not distributed in a plurality of regions around the world.

Disclosure of Invention

The invention aims to solve the technical problem of providing a low-orbit satellite (full) autonomous orbit and attitude determination method based on non-navigation satellite signals, which is suitable for integrated orbit and attitude determination of low-orbit satellites, can realize real-time high-precision orbit and attitude determination, improves the self-adaption and self-learning of the orbit and attitude determination of the low-orbit satellites, and gradually establishes, updates and perfects a global geographic feature point library, namely the autonomous orbit and attitude determination method based on GNSS/INS/SS/SLAM/SMNS or INS/SS/SLAM/SMNS.

Aiming at the problems that the current low-earth-orbit satellite orbit and attitude determination is excessively dependent on a GNSS satellite and a ground monitoring station and the observation data of a globally uniformly distributed ground tracking station cannot be used, the effective technical method can reduce the dependence of the low-earth-orbit satellite on the navigation satellite and the ground monitoring station, can meet the low-earth-orbit attitude determination requirements of the GNSS satellite in the extreme periods such as failure and the like including a Beidou system, provides redundancy and check for the orbit determination of the satellite-borne GNSS and the ground monitoring station in the ordinary period, and can gradually form and enrich and distribute a global geographic feature point library.

The method of the invention does not depend on satellite-borne GNSS and ground and satellite-borne orbit determination equipment such as SLR, DORIS, PRARE and the like (namely, an artificial signal emission source is not needed to continuously emit electromagnetic signals), and adopts a low orbit satellite autonomous orbit and attitude determination method based on INS/SS/SMNS/SLAM combined technology. The method can be redundantly implemented with non-autonomous orbit determination systems such as SLR and GNSS, can also be implemented independently, embodies autonomy, ensures precision, has good response capability particularly to abnormal states when the non-autonomous orbit determination systems fail, and has great practical significance for national political, economic and military safety. Different from the excessive dependence on navigation satellite signals in the past, the navigation satellite and other electromagnetic signals are very easily interfered, so that the space network node dependent on the navigation satellite signals is used for orbit determination, although the high precision can be achieved, the hidden danger exists, and the autonomous orbit and attitude determination work of the space nodes is indispensable due to the national safety factors such as politics, economy, military and the like of the aerospace work. Therefore, the independent orbit and attitude determination system of the space node which can operate autonomously and can learn by itself is established under the condition of the current limited method and is of great importance.

The method utilizes the existing calibration field in China, when the GNSS fails in an emergency and the satellite passes through the space above China in the daytime, the INS/SS navigation error can be corrected by utilizing the characteristic points in the image, so that the absolute accuracy of the system is improved, and the system is semi-autonomous orbit and attitude determination; when the satellite flies over the non-Chinese local area without ground control points, the INS/SS/SLAM method is used for keeping the precision, and the full-automatic orbit and attitude determination is carried out at the moment. When the GNSS is available in a normal period, the GNSS can be used, and the GNSS/INS/SS or INS/SS navigation error can be corrected by using the characteristic points in the image when the satellite passes the space above China in the daytime, so that the absolute accuracy of the system is improved; when the satellite flies over the non-Chinese local area without ground control points, the GNSS/INS/SS/SLAM or INS/SS/SLAM method is used for keeping the precision. The time period of the scheme is integrated, so that the full arc section orbit and attitude determination is realized.

Step S1, the integrated calibration of the autonomous orbit and attitude determination system INS/SS/SLAM/SMNS or GNSS/INS/SS/SLAM/SMNS is realized, which specifically comprises the following substeps;

step S11, establishing an error equation for the on-board GNSS receiver, the satellite-borne inertial sensor INS, the satellite-borne five-lens camera Cameras and the star sensor SS, and performing error calibration of the devices;

step S12, realizing the online calibration of the geometric relationship between the camera parameters and the multiple cameras;

step S13, synchronizing the exposure time of the camera and periodically checking;

step S14, establishing a navigation center mechanism, analyzing the position and attitude relation among the GNSS, the Cameras, the SS, the carrier mass center and the INS by taking the INS as a navigation center, and uniformly converting the positions and the attitude relation into a navigation center coordinate system;

and step S2, the system realizes orbit and attitude determination and simultaneously forms a geographic feature point library, continuously updates and perfects the global geographic feature point library, namely, the global geographic feature point library is formed while measuring, so that the high-precision time-phase invariant geographic feature point library is continuously updated. Aiming at the early stage, the middle stage and the later stage of the operation of the autonomous orbit and attitude determination system, the operation flow of the system is divided into the following four conditions: 1) in normal period, there is a control point area; 2) in an emergency period, a control point area exists; 3) in normal period, no control point area; 4) in an emergency period, no control point area exists; aiming at different conditions, different sensor devices in the system are combined to solve the high-precision spatial position and attitude of the satellite;

and step S3, the low orbit satellite real-time high-precision integrated orbit and attitude determination is realized by using the autonomous orbit and attitude determination system based on the non-navigation satellite signals.

Further, the specific implementation manner of step S12 is as follows,

the single-camera calibration model is represented as:

△x＝(x-x0)(k₁r²+k₂r⁴+k₃r⁶)+p₁(r²+2(x-x0)²)+2p₂(x-x0)(y-y₀)-b₁(x-x₀)+b₂(y-y₀)

△y＝(y-y₀)(k₁r²+k₂r⁴+k₃r⁶)+2p₁(x-x₀)(y-y₀)+p₂(r²+2(y-y₀)²)+b₁(y-y₀)

wherein r is the distance from the image point to the image main point, and x and y are the coordinates of the image point; Δ x, Δ y are coordinate corrections; x_S，Y_S，Z_SIs a coordinate of a camera station, X_A，Y_A，Z_AIs the coordinate of an object point; x is the number of₀，y₀Is a principal point; f is the main distance of the camera; k1, k2 and k3 are radial distortion parameters; p1 and p2 are eccentric distortion parameters; b1 and b2 are distortion parameters in an area array;

establishing a geometric relation between the camera and the INS centroid through a self-checking bundle adjustment equation with relative attitude parameters:

wherein, (X, Y, Z) is the coordinates of an object point; (x)_A,y_A,-f_A) The coordinates of image points in an image space coordinate system of the side-view camera are (u, v, w) the coordinates of the INS center in an image space auxiliary coordinate system, (X)_GNSS,Y_GNSS,Z_GNSS) As phase centre coordinates, R, of the satellite receiver antenna_INSTo represent

Constructing an orthogonal transformation matrix; r_misRepresenting a transformation matrix from the image space coordinate system to the INS coordinate system; r_ΔA rotation matrix obtained for angular elements in the relative poses of the side view camera and the intermediate vertical camera; (X)_Δ,Y_Δ,Z_Δ) Is a relative pose line element of the side view camera and the intermediate vertical camera; (X)_INS,Y_INS,Z_INS) Representing the INS measured camera initial line element values; (X)₀,Y₀,Z₀)、(X₁,Y₁,Z₁) Representing the error correction parameters of the INS camera line element coordinate drift system;

representing the initial angular element value of the shooting station of inertial navigation measurement;

the system drift correction parameter represents the change of the INS system camera angle element along with the time t, and the t represents the time for the INS to acquire the camera coordinate; t is t₀Represents a reference time; λ represents a proportionality coefficient;

order to

R′_ΔConverting the angular postures of the image space coordinate system and the INS coordinate system of each vision camera into a matrix; (X'_Δ,Y′_Δ,Z′_Δ) Line elements between the image space coordinate system and the INS coordinate system of each vision camera;

and (4) solving the correction number of the coordinates of the encrypted point and the correction numbers of other parameters through a normal equation of the adjustment model, and finally calculating the position and the posture of the image and the object space coordinates of the encrypted point.

Further, in step S2, for the high-precision time-phase-invariant geographic feature point library, the time-phase-invariant feature points are extracted, and due to the descriptors of the ground salient feature points, the gradient histograms of textures near the feature points are used, however, as time goes on and seasons change, the gradient strength and even direction of texture of some ground objects change, which may result in matching failure. Extracting time-phase invariant feature points, namely extracting and matching features through images of the same region in different periods, classifying the successfully matched feature points and mismatched feature points in different periods through a machine learning method, establishing a time-phase invariant feature point training library, and removing non-time-phase invariant feature points from the extracted feature points, wherein the phase comprises a phase of establishing a ground significant feature point library and a phase of positioning images. On the other hand, a descriptor insensitive to gradient strength and direction (left and right direction) is adopted, and segmented Euclidean distance is designed to calculate the similarity of the descriptor, so that the influence caused by strength change is relieved; planning the gradient direction from 0-360 degrees to 0-180 degrees, and removing the mismatch problem caused by the change of the gradient direction (left and right).

Further, in step S2, for 1) the normal period, there is a control point region, and the high-precision spatial position and attitude of the satellite are solved in four cases: 1a) when the ground control points are more (the number of the control points on a single image is more than M)₁Secondly), adopting a single-chip backward intersection to assist the GNSS/INS/SS or the INS/SS to carry out integrated orbit and attitude determination; 1b) when the ground control point is less than M₂If the number of the matched feature points is more (the number of the matched feature points on two continuous images of the camera with the same visual angle is more than M)₃Thirdly), adopting GNSS/INS/SS or INS/SS auxiliary beam method for block adjustment; 1c) when the ground control point is less than M₂If the number of the matched feature points is less (the number of the matched feature points on two continuous images of the camera with the same view angle is less than M)₄Thirdly), adopting GNSS/INS/SS or INS/SS to assist the adjustment of the area network with the beam method with relative attitude parameters; 1d) when the ground control point is less than M₂If the matched feature points are rare (the total number of the matched feature points on the 5 images of the same camera station is less than M)₅And secondly), the INS difference value of short time for positioning and attitude determination by combining INS-assisted continuous sequential images based on the GNSS/INS/SS or INS/SS is used as the true value of the relative attitude of the images, N images are fixedly connected and then matched, and then the adjustment of the area network by the GNSS/INS/SS or INS/SS assisted beam method is carried out.

Further, step S2 is directed to 2) an emergency period (GNSS is not available), there is a control point region, and the high-precision spatial position and attitude of the satellite are solved in the following four cases: 2a) when the ground control points are more (the number of the control points on a single image is more than M)₁Respectively), adopting a single-chip rear intersection auxiliary INS/SS to carry out integrated orbit and attitude determination; 2b) when the ground control point is less than M₂If the number of the matched feature points is more (the number of the matched feature points on two continuous images of the camera with the same visual angle is more than M)₃Thirdly), adopting INS/SS auxiliary beam method block adjustment; 2c) when the ground control point is less than M₂If there are few matching feature points (same view angle camera)The number of the matched characteristic points on the two continuous images is less than M₄Firstly), adopting INS/SS to assist adjustment of the area network with the beam method of relative attitude parameters; 2d) when the ground control point is less than M₂If the matched feature points are rare (the total number of the matched feature points on the 5 images of the same camera station is less than M)₅And secondly), adopting INS-assisted continuous sequential images based on INS/SS to jointly position and fix the attitude, namely using the INS difference value of a short time as a true value of the relative attitude of the images, matching after N images are fixedly connected with N images, and then performing INS/SS assisted beam method block adjustment.

Further, in step S2, for 3) in the normal period, there is no control point region, and the high-precision spatial position and attitude of the satellite are solved in two cases: 3a) when the SLAM is available, adopting GNSS/INS/SS/SLAM or INS/SS/SLAM to carry out integrated orbit and attitude determination and establishing a global geographic feature point library and continuously updating the global geographic feature point library; 3b) when the matched feature points are rare (the total number of the matched feature points on the 5 images of the same camera station is less than M)₆And (c) when SLAM is unavailable, using an INS assisted continuous sequential image joint positioning and attitude determination based on GNSS/INS/SS or INS/SS (within a short period of time, relative positions and relative attitudes between images are considered known, forming "temporally continuous spatially separated images").

Further, in step S2, for 4), in the emergency period, there is no control point region, the high-precision spatial position and attitude of the satellite are solved in two cases: 4a) when the SLAM is available, adopting INS/SS/SLAM to carry out integrated orbit and attitude determination and establishing a global geographic feature point library and continuously updating the global geographic feature point library; 4b) when the matched feature points are rare (the total number of the matched feature points on the 5 images of the same camera station is less than M)₆And when the SLAM is unavailable, the INS based on the INS/SS is adopted to assist continuous sequential images to jointly position and fix the posture.

Furthermore, the specific implementation mode of adopting the INS/SS to carry out integrated orbit and attitude determination is as follows,

establishing a state equation of the INS/SS combined navigation system, comprising the following steps: attitude error equation, velocity error equation and position error equation, the state variables of which include the misalignment angle of the mathematical platform, the velocity error, the position error, the gyro constant drift of the inertial device and the accelerometer zero offset; establishing an INS error model by taking a local geographic coordinate system ENU as a navigation coordinate system, wherein a system state equation is as follows:

the state variables are as follows:

wherein (phi)_E φ_N φ_U) Representing the northeast ENU attitude error (delta V) of the satellite-borne inertial sensor INS in the three-axis direction_EδV_N δV_U) Representing the speed error, (delta L delta lambda delta h) representing the position error of the satellite-borne inertial sensor INS, (epsilon)_E ε_N ε_U) Representing the gyro constant drift of the inertial device,

indicating the zero offset of an accelerometer of the inertial sensor, (delta Ax delta Ay delta Az) indicating the installation error of the star sensor; matrix array

The transformation matrix from the carrier coordinate system to the navigation coordinate system is expressed as a matrix F, a coefficient matrix and a matrix

W in W_εx，w_εy，w_εzAnd

zero mean random white noise for the gyroscope and accelerometer, respectively;

defining a transformation matrix from the carrier coordinate system to the navigation coordinate system under the navigation coordinate system as

The transformation matrix from the navigation coordinate system to the mathematical platform coordinate system is

The transformation matrix from the carrier coordinate system to the mathematical platform coordinate system is

The relation among the carrier coordinate system, the navigation coordinate system and the mathematical platform coordinate system is as follows:

the inertial navigation system and the star sensor both output attitude angles from a carrier to a navigation coordinate system, wherein an attitude matrix output by the star sensor is an attitude matrix from a converted carrier coordinate system to the navigation coordinate system, and an attitude matrix output by the inertial navigation system is a conversion matrix from the carrier coordinate system to a recursion mathematical platform calculation coordinate system; defining INS to obtain pitch angle P through calculation of strapdown inertial navigation mathematical platform₁Transverse roll angle R₁Heading angle Y₁The star sensor measures the pitch angle P₂Transverse roll angle R₂Heading angle Y₂Subtracting the three-axis attitude error angle by the difference between the three attitude error angles:

converting the three-axis error attitude angle into a mathematical platform coordinate system to obtain a misalignment angle attitude matrix, and taking the attitude information difference value measured by the attitude output and star sensor of the inertial navigation system as the observed quantity of a measurement equation, so that an observation equation Z is obtained₁＝H₁X+V₁In particular toExpressed as:

wherein [ w_x w_y w_z]^TFor measuring zero mean Gaussian white noise of the noise V, measuring the matrix

The elements in (1) are obtained by calculation from attitude information measured and output by the star sensor.

Further, M₁Has a value of 4, M₂Is taken as 3, M₃Is 12 to 15, M₄Is 12 to 15, M₅Is 5 to 10, M₆The value of (1) is 5-10.

According to the technical scheme, the method has the following advantages:

in step S1, the problems of resolution and breadth are solved by carrying the satellite-borne five cameras, and meanwhile, the establishment of a high-precision autonomous orbit and attitude determination system is realized by utilizing an in-orbit periodic calibration technology.

In step S1, a navigation center mechanism is adopted, so that any sensor can be used as a navigation center and the center of mass of the carrier can be unified.

In step S2, the low-orbit satellite integrated autonomous orbit and attitude determination method and technique are also applicable to medium and high-orbit satellites, and can realize related parameter setting based on different precisions, thereby further realizing the multi-mode full-orbit integrated orbit and attitude determination system.

In step S2, the INS is used as a main line for tight coupling to realize INS-assisted star atlas identification, INS-assisted SMNS, INS-assisted SLAM, and the like.

The invention has the beneficial effects that: the invention relates to a low-orbit satellite autonomous orbit and attitude determination method based on non-navigation satellite signals, which realizes the real-time high-precision integrated autonomous orbit and attitude determination of a low-orbit satellite by acquiring image data, global positioning data, inertial navigation data and star sensor data, and gradually establishes a global high-precision time phase geographic feature point library, namely realizes the INS/SS/SLAM/SMNS orbit and attitude determination based on the non-navigation satellite signals. Compared with the traditional low-orbit satellite orbit and attitude determination method, the infeasibility of global arrangement of ground control points is avoided, and meanwhile, the method is suitable for normal operation of the orbit and attitude determination system under the condition of abnormal satellites in an extraordinary period. Particularly, the establishment of a global high-precision time phase invariant geographic feature point library can realize INS/SS/SMNS high-precision orbit and attitude determination, has important strategic significance, and greatly improves the safety of a spatial information network in an emergency.

The integrated autonomous orbit and attitude determination system based on the non-navigation satellite signals is realized, and important guarantee is provided for the fields of national security and military. Meanwhile, important technical approaches and exploration routes are provided for further realizing the tamping foundation of the multi-mode full-track integrated orbit and attitude determination system.

Drawings

FIG. 1 is a prior operational diagram of an embodiment of the present invention;

FIG. 2 is a diagram of mid-term operation of an embodiment of the present invention;

FIG. 3 is a later operational view of an embodiment of the present invention;

fig. 4 is a flowchart illustrating the overall operation of the embodiment of the present invention.

Detailed Description

The principles and features of this invention are described below in conjunction with the following drawings, which are set forth by way of illustration only and are not intended to limit the scope of the invention.

The method comprises the steps of firstly, distributing a certain number of characteristic points on the ground, measuring geographical coordinates of the characteristic points in advance by a certain method, and then storing images of the characteristic points and the geographical coordinates of the characteristic points in a satellite-borne processing center; and meanwhile, a fixed star distribution map is drawn by a certain method and stored in a satellite-borne processing center to form a navigation star reference library. A star sensor and a remote sensing camera are installed on a low-orbit satellite, a star map is captured by the star sensor camera, star point extraction and star point coordinate calculation are carried out, image points formed by star bodies are matched with a navigation star library, position coordinates of the star bodies in a celestial coordinate system, which correspond to the image points, can be obtained through analysis, and finally a carrier attitude observation equation is established; meanwhile, the camera shoots a ground image, the feature points in the real-time shot image are matched with the feature points in the reference image through a certain matching algorithm, and then the position and the posture of INS navigation, the coordinates of the image points of the feature points and the known geographic coordinates are used as observation information, and a collinear equation is established through a rear intersection method. The system designs a Kalman filter, an attitude observation equation and a collinear equation are used as an observation equation of the Kalman filter, an INS error equation is used as a system state equation, and finally the accurate position and attitude of the satellite, and device errors of the INS, the SS and a photographic camera (satellite image) are solved through the Kalman filter. The combination technology mainly comprises a star map rapid identification and a rapid and accurate matching algorithm of image characteristic points. The INS navigation information is used for assisting the star map identification of the SS, so that the success rate and the measurement accuracy are improved; the image matching relates to preprocessing such as matching region selection, image enhancement, refinement, denoising and geometric distortion correction, feature extraction, image analysis and evaluation, a matching algorithm and the like, and is used for avoiding gross error and precision reduction caused by misregistration. Therefore, the scheme also uses the INS navigation information to assist geometric correction of the image and improve the image distortion resistance of image matching on one hand, and helps the image matching to narrow the search range on the reference image on the other hand, so that the matching accuracy is improved, the matching calculation amount is greatly reduced, and the real-time performance of the image matching is improved.

In areas without known ground coordinate feature points, the relative position and the attitude of the LEO satellite are estimated by using an SLAM method, and combined navigation calculation is carried out with the INS/SS. SLAM firstly uses the stereo image and the initial navigation result information of INS to carry out stereo measurement on some characteristic points to obtain geographic coordinates, and then uses the spatial information of the characteristic points and the INS/SS navigation information to carry out filtering in subsequent images to obtain the accurate new position and attitude of the satellite and the device error of the sensor. In the method of the present invention, the feature point image matching work is substantially the same as the image matching of the feature points with known coordinates, and is not described herein again, nor is it necessary to distinguish between normal and abnormal periods.

According to the scheme, the on-line calibration of the multiple cameras is realized through a satellite-borne GNSS/INS/SS combined positioning and attitude determination system, instead of pre-calibration on the ground, the ground calibration only provides a technical route of an initial value. And the INS is used as a navigation center to establish the posture and position relations between the rest sensors, the carrier and the INS. In a usual period, the satellite-borne GNSS can normally work, and the system can finish calibration work by using the period, so that in an emergency period when the GNSS cannot work due to suppression and the like, the system can realize autonomous orbit and attitude determination by using the INS/SS/SMNS combination.

The method has the important characteristics that a global feature point library is gradually established through INS/SS/SLAM combination, unknown feature points are gradually changed into known feature points, and the geographic coordinate accuracy of the feature points is gradually improved. On the basis that a camera is calibrated, satellite remote sensing mapping can be carried out through direct geographic reference of satellite-borne INS/SS, so that global mapping is carried out by utilizing the characteristic that a low-orbit satellite flight path is in cyclic reciprocation during the working period of the INS/SS/SLAM, a feature point library which is distributed all over the world is built step by step, and in the normal working period of the satellite-borne GNSS, GNSS/INS/SS combined positioning and attitude determination can be utilized to improve the precision of a remote sensing platform. When the coordinates of the feature points are unknown, namely the INS/SS/SLAM algorithm is used for solving the front and back displacement and the attitude change of the satellite, the INS/SS/SLAM algorithm gradually evolves into the INS/SS/SMNS algorithm along with more and more geographical coordinates of the feature points, and the high-precision spatial position and attitude of the satellite can be solved. Meanwhile, in daily orbit and attitude determination work, the accuracy optimization of the geographic coordinates of the feature points is a continuous process, namely, the remote sensing images at the later stage are used for orbit and attitude determination on one hand and solving the coordinates of the feature points on the other hand to optimize the accuracy of the original coordinates of the feature points.

The technical scheme for solving the technical problems is as follows: a low orbit satellite autonomous orbit and attitude determination method based on non-navigation satellite signals comprises the following steps:

step S1, an error equation model is needed to be built for four devices of an on-board GNSS receiver, an on-board inertial sensor INS, an on-board five-camera device and a star sensor SS to weaken the influence of errors and carry out self calibration of the devices, a navigation center mechanism is built in the calibration process of the geometric relationship of a synchronous camera set (the geometric relationship between camera parameters and multiple cameras), the geometric relationship between a camera and the center of mass of a low-orbit satellite and the exposure synchronization of the camera, meanwhile, the position and attitude relationship between the INS and each sensor and carrier acquisition is analyzed in the integrated calibration process of an autonomous orbit determination and attitude determination system INS/SS/SLAM/SMNS or GNSS/INS/SS/SLAM/SMNS, and a technical route that the on-board GNSS system periodically carries out on-orbit calibration and the ground calibration only provides initial values is adopted. The specific implementation way of step S1 is:

step S11, establishing an error equation for a satellite-borne inertial sensor INS, a satellite sensor SS, an on-satellite GNSS receiver and a satellite-borne five-camera device (a single camera establishes an error model and checks the error) and performing error checking on the device (the technology is mature and no surplus exists);

step S12, firstly, checking a single camera in the satellite-borne five-camera device, and then establishing a geometric relationship between the single camera and the INS (inertial navigation System) mass center;

the single-camera calibration model is represented as:

△x＝(x-x₀)(k₁r²+k₂r⁴+k₃r⁶)+p₁(r²+2(x-x₀)²)+2p₂(x-x₀)(y-y₀)-b₁(x-x₀)+b₂(y-y₀)

wherein r is the distance from the image point to the image main point, and x and y are the coordinates of the image point; Δ x, Δ y are coordinate corrections; x_S，Y_S，Z_SIs a coordinate of a camera station, X_A，Y_A，Z_AIs the coordinate of an object point; x is the number of₀，y₀Is the coordinate of the image principal point; f is the main distance of the camera; k1, k2 and k3 are radial distortion parameters; p1 and p2 are eccentric distortion parameters; b1 and b2 are distortion parameters in an area array.

inertial navigation measurementDetermining an initial angle element value of the camera station;

the system drift correction parameter represents the change of the INS system camera angle element along with the time t, and the t represents the time for the INS to acquire the camera coordinate; t is t₀Represents a reference time; λ represents a scaling factor.

Order to

R′_ΔConverting the angular postures of the image space coordinate system and the INS coordinate system of each vision camera into a matrix; (X'_Δ,Y′_Δ,Z′_Δ) Is the line element between the spatial coordinate system of each view camera and the INS coordinate system.

The correction number of the coordinates of the encrypted point (characteristic object point) and the correction numbers of other parameters are obtained through the normal equation of the adjustment model, and finally the position and the posture of the image and the object space coordinates of the encrypted point are calculated.

In step S12, for the satellite image back-meeting, the width determines the system structural strength and the resolution determines the system view measurement accuracy level. In order to ensure the image resolution while enlarging the image frame, a scheme for absolutely positioning and positioning a plurality of synchronous high-resolution images is provided, namely, a group of five high-resolution cameras is arranged on a low-orbit satellite, the image size of each camera is very small, only a few kilometers are needed (the ground coverage designed by the scheme is about 4km multiplied by 4km), but the resolution is high and can reach 20cm, the photographic optical axis of each camera is spread outwards, the geometric relationship between the cameras can be determined by first calibration at the image point on the ground of 100km, so that under the condition of synchronous shooting of a plurality of cameras, then, one image (covering the ground area about 100km x 100km) with only the four corners and the central area having images and the other areas being blank is formed, and by matching the feature points in the four corners and the central photographing area, the absolute position and attitude of the low earth orbit satellite can be determined by an image matching rear intersection method. The mode of matching a plurality of synchronous high-resolution images can simultaneously realize high structural strength and high observation precision of the system.

And step S13, synchronizing the exposure time of the camera and periodically checking, wherein the exposure synchronization of the camera belongs to the hardware design content, a uniform time signal source is adopted to control the synchronous exposure of the multiple cameras, and because the synchronous precision of the pulse signal is nanosecond and the synchronous precision of the shutter of the camera is millisecond and is difficult to be higher, the synchronous error exists. However, this synchronization error will be absorbed by the system during camera calibration, and its residual will mainly come from the stability of the exposure delay of the camera, so long as the exposure of the camera is not fast or slow, the precision can be guaranteed. However, because of long period drift of exposure delay of the camera, the camera still needs to be periodically checked and corrected during system operation;

and step S14, the autonomous orbit and attitude determination system is used for integrated calibration, and a technical route that the satellite-borne system is used for on-orbit calibration and the ground calibration only provides initial values is adopted for the geometric relations among the five photographic cameras, the geometric relations among the cameras and the INS, the cameras and the SS centroid and the satellite centroid. And establishing a navigation center mechanism, analyzing the position and posture relation between other sensors and carriers and the INS by taking the INS as a navigation center, and uniformly converting the positions and the posture relation into a navigation center coordinate system. In a usual period, the satellite-borne GNSS can normally work, and the system can finish calibration work by using the period, so that the system can realize fully autonomous orbit and attitude determination by using the INS/SS/SMNS combination in an emergency period when the GNSS cannot work due to suppression and the like.

The geometric relationship between the camera and the SS centroid, the geometric relationship between the camera and the satellite centroid is similar to the geometric relationship between the camera and the INS centroid, but need not be so expressed. The method adopts a navigation center mechanism, any sensor can be used as a navigation center and establishes the relation with the camera, and the geometric relation between the camera and the INS centroid is established as the above formula. The INS is used as a navigation center, and due to the limitation of processing precision, the installation error between the satellite sensor and the inertial navigation system can often reach angular classification. Factors such as the age and the parallax cause slight changes of the star direction, so that the star sensor measurement errors are caused, but the errors caused by the factors are in the order of attosecond. In contrast, the satellite-sensitive mounting error becomes one of the main factors limiting the satellite-sensitive measurement accuracy. Therefore, the position and attitude information of all the sensors is transformed to a navigation center mainly based on INS in a unified planning mode, and a relation with the carrier is established, namely a rotation matrix of the three-axis direction of the center of mass of the carrier and the three-axis direction of the navigation center. The attitude information, the position information and the satellite-sensitive measurement information output by the satellite-borne inertial sensor INS are used for constructing observed quantities, a Kalman filtering model is established, and calibration compensation of a satellite-sensitive installation error angle is realized. Specifically, see the INS/SS filter formula, the observation equation or the measurement equation includes three error sources: and obtaining the installation error of the star sensor and calibrating and compensating the installation error of the star sensor by the attitude error of the star inertial sensor INS, the position error of the star inertial sensor INS and the installation error of the star sensor.

Considering the installation error of the star sensor, the installation error delta A of the star sensor along the three directions of the carrier x, y and z can be realized_i(i ═ x, y, z) is considered to be a random constant, i.e.

Taking a local geographic coordinate system (ENU) as a navigation coordinate system, the INS/SS system state equation is as follows:

is the state vector derivative, where the state variable is represented as:

W in W_εx，w_εy，w_εzAnd

zero mean random white noise for the gyro and accelerometer, respectively.

And step S2, when the geographic feature point coordinates are unknown, the GNSS/INS/SS/SLAM or INS/SS/SLAM algorithm is used for solving the front and back displacement and the attitude change of the satellite, and as the geographic feature point coordinates of the feature points are solved more and more to enrich the global high-precision time phase-invariant geographic feature point library, the INS/SS/SLAM algorithm gradually evolves into the INS/SS/SMNS algorithm to solve the high-precision spatial position and attitude of the satellite. In the early stage of system operation, a corresponding orbit and attitude determination method is adopted according to the number of control points and whether a navigation satellite signal is available, a geographic feature point library is preliminarily formed, and the global geographic feature point library is continuously updated and perfected; in the middle stage of system operation, a corresponding orbit and attitude determination method is adopted according to the number of control points and whether a navigation satellite signal is available, a high-precision time-phase invariant geographic feature point library is gradually established on the basis of the previous stage operation, the global coverage of the point library is continuously expanded, and the accuracy of the point library is continuously improved; in the later stage of system operation, a corresponding orbit and attitude determination method is adopted according to the number of control points and whether navigation satellite signals are available, and in the later stage of system operation, a high-precision time-phase-invariant global geographic feature point library is basically established and perfected and can be distributed over seven continents around the world and islands in the sea; the specific implementation way of step S2 is:

step S21, in the earlier stage of system operation, the system initially forms a geographic feature point library while realizing orbit and attitude determination, continuously updates and perfects the global geographic feature point library, namely, the global geographic feature point library is formed while measuring, and the system in the earlier stage of system operation contributes less to the establishment of the time phase invariant feature point library. The system operation flow is divided into four conditions: 1) in normal period, there is a control point area; 2) in an emergency period, a control point area exists; 3) in normal period, no control point area; 4) in an emergency, there is no control point area. In the normal period, there are four cases of control point region: 1a) when the ground control points are more (the number of the control points on a single image is more than M)₁Secondly), adopting a single-chip backward intersection to assist the GNSS/INS/SS or the INS/SS to carry out integrated orbit and attitude determination; 1b) when the ground control point is less than M₂If the number of the matched feature points is more (the number of the matched feature points on two continuous images of the camera with the same visual angle is more than M)₃Thirdly), adopting GNSS/INS/SS or INS/SS auxiliary beam method for block adjustment; 1c) when the ground control point is less than M₂If the number of the matched feature points is less (the number of the matched feature points on two continuous images of the camera with the same view angle is less than M)₄Thirdly), adopting GNSS/INS/SS or INS/SS to assist the adjustment of the area network with the beam method with relative attitude parameters; 1d) when the ground control point is less than M₂If the matched feature points are rare (the total number of the matched feature points on the 5 images of the same camera station is less than M)₅Secondly), the INS difference value of the short time for positioning and attitude determination by combining INS auxiliary continuous sequential images based on GNSS/INS/SS or INS/SS is taken as the true value of the relative attitude of the images, N images are fixedly connected and then matched, and then the adjustment of the area network by the GNSS/INS/SS or INS/SS auxiliary beam method is carried out;

in an emergency (GNSS not available) there is a control point zoneThe domains are divided into four cases: 2a) when the ground control points are more (the number of the control points on a single image is more than M)₁Respectively), adopting a single-chip rear intersection auxiliary INS/SS to carry out integrated orbit and attitude determination; 2b) when the ground control point is less than M₂If the number of the matched feature points is more (the number of the matched feature points on two continuous images of the camera with the same visual angle is more than M)₃Thirdly), adopting INS/SS auxiliary beam method block adjustment; 2c) when the ground control point is less than M₂If the number of the matched feature points is less (the number of the matched feature points on two continuous images of the camera with the same view angle is less than M)₄Firstly), adopting INS/SS to assist adjustment of the area network with the beam method of relative attitude parameters; 2d) when the ground control point is less than M₂If the matched feature points are rare (the total number of the matched feature points on the 5 images of the same camera station is less than M)₅Secondly), adopting INS-assisted continuous sequential images based on INS/SS to jointly position and fix the attitude, namely using INS difference values in a short time as true values of relative attitude of the images, matching after N images are fixedly connected with N images, and then performing INS/SS assisted beam method block adjustment;

in the normal period, the no-control-point region is divided into two cases: 3a) when the SLAM is available, adopting GNSS/INS/SS/SLAM or INS/SS/SLAM to carry out integrated orbit and attitude determination and establishing a global geographic feature point library and continuously updating the global geographic feature point library; 3b) when the matched feature points are rare (the total number of the matched feature points on the 5 images of the same camera station is less than M)₆When SLAM is unavailable, the INS based on GNSS/INS/SS or INS/SS is adopted to assist continuous sequential images to jointly position and fix the attitude (within a short period, the relative position and relative attitude between the images are known to form a 'time continuous space separation image');

in the emergency, the area without control point is divided into two cases: 4a) when the SLAM is available, adopting INS/SS/SLAM to carry out integrated orbit and attitude determination and establishing a global geographic feature point library and continuously updating the global geographic feature point library; 4b) when the matched feature points are rare (the total number of the matched feature points on the 5 images of the same camera station is less than M)₆Respectively) when the SLAM is unavailable, adopting INS based on INS/SS to assist continuous sequential images to jointly position and fix the posture;

as shown in fig. 1, the system is operated in the early stage: a1. the system runs for the first circle, the system is represented by a black line, the precision of the measured geographic characteristic point is continuously attenuated from high to low along with the running of the designed running track of the satellite until the lowest precision point is reached, the precision is improved when the system is close to a control point region arranged in China, namely the lowest precision point is not the last closed point in the circle;

b1. the system runs for the second circle, the grey line is adopted for representing, the satellite running tracks are the same (the outer circle is drawn for showing different effects, the tracks are actually the same), on the basis of the first circle, the global geographic feature point library area is updated, the updated area is close to the vicinity of the existing control point area, the precision of the measured geographic feature point is continuously attenuated from high to low until the lowest precision is reached, the precision is improved when the global geographic feature point library area is close, and at the moment, the lowest precision is pushed forwards;

c1. the process is circulated, so that the area of the existing global geographic feature point library is continuously expanded, and the lowest precision point is continuously pushed forward … …

Step S22, in the middle stage of system operation, the system operation flow is the same as the four conditions in the early stage of system operation (no redundancy), and at this time, a high-precision time-phase-invariant geographic feature point library is gradually built on the basis of early stage operation, the global coverage of the point library is continuously expanded, and the precision of the point library is continuously improved. The time phase invariant feature points are extracted, gradient histograms of textures near the feature points are adopted due to descriptors of the ground significant feature points, and however, the matching failure can be caused by the change of the gradient strength and even the direction of the textures of some ground features along with the change of time and seasons. Extracting time-phase invariant feature points, namely extracting and matching features through images of the same region in different periods, classifying the successfully matched feature points and mismatched feature points in different periods through a machine learning method, establishing a time-phase invariant feature point training library, and removing non-time-phase invariant feature points from the extracted feature points, wherein the phase comprises a phase of establishing a ground significant feature point library and a phase of positioning images. By adopting a descriptor insensitive to gradient strength and direction (left and right direction), designing segmented Euclidean distance to calculate the similarity of the descriptor, and slowing down the influence brought by strength change; planning the gradient direction from 0-360 degrees to 0-180 degrees, and removing the mismatch problem caused by the change of the gradient direction (left and right).

Fig. 2 shows the middle stage of the system operation: a2. after the system operation reaches the middle stage, about half of the global geographic feature point library is established, the precision of the point library is continuously improved, the attenuation is reduced from high to low, and the lower area is continuously reduced;

b2. the process is circulated, so that the area of the existing global geographic feature point library is continuously expanded, the lowest point of precision is continuously pushed forward, and the precision of the existing point library is continuously improved … …

Step S23, the system operation flow in the later stage of system operation is the same as the four conditions (no surplus) in the early stage and the middle stage of system operation, the system operation reaches the later stage, a global geographic feature point library which is unchanged when the system operation reaches the later stage is basically established and perfected, and the feature point library can be distributed over seven continents around the world and islands in the sea;

fig. 3 shows the system in the later operation stage: a3. in the later stage of system operation, establishing a high-precision time-phase-invariant global geographic feature point library which can be distributed over islands in seven continents and oceans in the world;

b3. the integrated autonomous orbit and attitude determination of the low orbit satellite based on the non-navigation satellite signal is basically realized.

The combined orbit and attitude determination method mentioned in step S2 is:

the method adopts the federal filtering theory and aims to provide a fault-tolerant integrated navigation system, takes INS with comprehensive navigation information, high output rate and absolutely guaranteed reliability as a public reference system, and combines the INS with other subsystems in pairs to form a plurality of sub-filters. All the sub-filters operate in parallel to obtain local optimal estimates established on the basis of local measurement of the sub-filters, the local optimal estimates are synthesized in the main filter according to a fusion algorithm to obtain global estimates established on the basis of all the measurement, so that the reliability of the whole system is improved, and a global suboptimal result with high autonomy and high reliability is obtained.

Compared with a loosely-coupled INS/SS/SMNS combination mode, on one hand, the SMNS in the tightly-coupled system can make full use of auxiliary information such as position and attitude provided by the INS to perform distortion correction on a real-time image and improve registration accuracy and anti-interference capability, and on the other hand, the SMNS position and speed measurement result can be directly sent to a filter to correct INS accumulated errors.

Based on an error model of the inertial navigation system, a state equation of the INS/SS integrated navigation system is established, wherein the state equation mainly comprises an attitude error equation, a speed error equation and a position error equation, and state variables of the INS/SS integrated navigation system comprise a mathematical platform misalignment angle, a speed error and a position error as well as gyro constant drift and accelerometer zero offset of an inertial device. Taking a local geographic coordinate system (ENU) as a navigation coordinate system, and establishing an INS error model, wherein a system state equation is as follows:

the state variables are as follows:

in the form of matrix

The matrix F is a coefficient matrix

W in W_εx，w_εy，w_εzAnd

zero mean random white noise for the gyro and accelerometer, respectively.

Carrier coordinate system toThe transformation matrix of the mathematical platform coordinate system is

the inertial navigation system and the star sensor can output attitude angles from a carrier to a navigation coordinate system, wherein an attitude matrix output by the star sensor is an attitude matrix from the converted carrier coordinate system to the navigation coordinate system, and an attitude matrix output by the inertial navigation system is a conversion matrix from the carrier coordinate system to a calculation coordinate system of a recursion mathematical platform. Defining INS to obtain pitch angle P through calculation of strapdown inertial navigation mathematical platform₁Transverse roll angle R₁Heading angle Y₁. Star sensor measuring pitch angle P₂Transverse roll angle R₂Heading angle Y₂. Subtracting the three-axis attitude error angle from the three-axis attitude error angle to obtain a three-axis attitude error angle:

and converting the three-axis error attitude angle into a mathematical platform coordinate system, thereby obtaining a misalignment angle attitude matrix. Taking the difference value of the attitude output of the inertial navigation system and the attitude information measured by the star sensor as the observed quantity of a measurement equation, and then observing an equation Z₁＝H₁X₁+V₁Can be specifically expressed as:

wherein [ w_x w_y w_z]^TFor measuring noiseZero mean Gaussian white noise of the sound V, the measurement matrix

The element (B) can be obtained by calculation from attitude information measured and output by the star sensor.

The INS/SMNS measurement equation can be expressed as:

in the formula (7), V₂Noise is measured for the image.

Then the INS/SS/SMNS combination system measurement equation is:

the GNSS/INS measurement equation is:

in the formula (9), the reaction mixture is,

respectively represent the velocity of the INS in the northeast direction, L_INS、λ_INS、H_INSRepresenting INS latitude, longitude and altitude, and similarly representing corresponding values of GNSS satellites by GNSS superscripts, wherein Rm is the meridian curvature radius; rn is the curvature radius of the unitary mortise ring.

In the normal period, GNSS/INS/SS/SMNS can be carried out to carry out orbit and attitude determination, and the measurement equation is as follows:

in conclusion, in the initial running stage of the satellite, GNSS/INS/SS/SLAM or INS/SS/SLAM can be adopted to fix orbit and attitude in the normal period, and the GNSS of the navigation satellite signal is not taken as an essential means; and in an emergency or in GNSS failure, the INS/SS/SLAM can still be used for orbit and attitude determination. And when the condition limit can not carry out SLAM, the INS/SS or the GNSS/INS/SS is adopted to carry out orbit determination and attitude determination (the combination mode is not redundant here).

After the satellite runs for a long time and a rich global geographic feature point library is established, GNSS/INS/SS/SMNS or INS/SS/SMNS can be adopted in a normal period, and INS/SS/SMNS or INS/SS can be adopted in an emergency period. When SMNS is unavailable, GNSS/INS/SS or INS/SS can be used in normal period, and INS/SS can be used in emergency period.

And step S3, forming a global high-precision time phase invariant geographic feature point library, and realizing the autonomous orbit and attitude determination based on the GNSS/INS/SS/SLAM/SMNS or INS/SS/SLAM/SMNS. The low-orbit satellite real-time high-precision integrated orbit and attitude determination can be completed by utilizing the INS/SS/SMNS high-precision fully-autonomous orbit and attitude determination technology based on non-navigation satellite signals.

The specific implementation way of step S3 is:

and step S31, gradually improving the geographic coordinate precision of the feature points, performing global mapping by using the characteristic of cyclic reciprocation of the flight path of the low-earth satellite during the SLAM working period, gradually enriching until a feature point library distributed all over the world is perfected, and improving the precision of the remote sensing platform by using GNSS/INS/SS combined positioning and attitude determination during the normal working period of the satellite-borne GNSS. Therefore, a global high-precision time phase invariant geographic feature point library is realized.

And step S32, the INS/SS/SLAM algorithm is finally evolved into the INS/SS/SMNS algorithm to solve the high-precision spatial position and attitude of the satellite. And forming a high-precision time phase invariant global geographic feature point library, and realizing the autonomous orbit and attitude determination method and system based on the GNSS/INS/SS/SLAM/SMNS or INS/SS/SLAM/SMNS. The method and the system for low-orbit satellite real-time high-precision integrated orbit and attitude determination can be completed by utilizing the INS/SS/SMNS high-precision fully-autonomous orbit and attitude determination technology based on non-navigation satellite signals.

The specific embodiments described herein are merely illustrative of the spirit of the invention. Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.

Claims

1. A low orbit satellite autonomous orbit and attitude determination method based on non-navigation satellite signals is characterized by comprising the following steps:

step S2, the system forms a geographic feature point library while realizing orbit and attitude determination, continuously updates and perfects the global geographic feature point library, namely, the global geographic feature point library is formed while measuring, so that the high-precision time phase-invariant geographic feature point library is continuously updated; aiming at the early stage, the middle stage and the later stage of the operation of the autonomous orbit and attitude determination system, the operation flow of the system is divided into the following four conditions: 1) in normal period, there is a control point area; 2) in an emergency period, a control point area exists; 3) in normal period, no control point area; 4) in an emergency period, no control point area exists; aiming at different conditions, different sensor devices in the system are combined to solve the high-precision spatial position and attitude of the satellite;

and step S3, completing low-orbit satellite real-time high-precision integrated orbit and attitude determination by using the autonomous orbit and attitude determination system based on the non-navigation satellite signals.

2. The method for autonomously determining the orbit and the attitude of the low orbit satellite based on the non-navigational satellite signals as claimed in claim 1, wherein: the specific implementation of step S12 is as follows,

the single-camera calibration model is represented as:

△x＝(x-x₀)(k₁r²+k₂r⁴+k₃r⁶)+p₁(r²+2(x-x₀)²)+2p₂(x-x₀)(y-y₀)-b′₁(x-x₀)+b′₂(y-y₀)

△y＝(y-y₀)(k₁r²+k₂r⁴+k₃r⁶)+2p₁(x-x₀)(y-y₀)+p₂(r²+2(y-y₀)²)+b′₁(y-y₀)

wherein r is the distance from the image point to the image main point, and x and y are the coordinates of the image point; Δ x, Δ y are coordinate corrections; x_S，Y_S，Z_SIs a coordinate of a camera station, X_A，Y_A，Z_AIs the coordinate of an object point; x is the number of₀，y₀Is a principal point; f is the main distance of the camera; k is a radical of₁、k₂、k₃Is a radial distortion parameter; p is a radical of₁、p₂Is an eccentric distortion parameter; b'₁、b′₂Is an in-plane distortion parameter;

Constructing an orthogonal transformation matrix; r_misRepresenting a transformation matrix from the image space coordinate system to the INS coordinate system; r_ΔA rotation matrix obtained for angular elements in the relative poses of the side view camera and the intermediate vertical camera; (X)_Δ,Y_Δ,Z_Δ) Is a relative pose line element of the side view camera and the intermediate vertical camera; (X)₀,Y₀,Z₀)、(X₁,Y₁,Z₁) Representing the error correction parameters of the INS camera line element coordinate drift system;

order to

R′_ΔFor each visual phaseA transformation matrix of the angular postures of the image space coordinate system and the INS coordinate system; (X'_Δ,Y′_Δ,Z′_Δ) Line elements between the image space coordinate system and the INS coordinate system of each vision camera;

and (3) solving the correction number of the coordinates of the encrypted point and the correction numbers of other parameters through a normal equation of the adjustment model, and finally calculating the position and the posture of the image and the object space coordinates of the encrypted point.

3. The method for autonomously determining the orbit and the attitude of the low orbit satellite based on the non-navigational satellite signals as claimed in claim 1, wherein: in step S2, for 1) the normal period, there is a control point region, and the high-precision spatial position and attitude of the satellite are solved in the following four cases: 1a) when there are more ground control points, i.e. the number of control points on a single image is greater than M₁Secondly, a single-chip backward intersection is adopted to assist the GNSS/INS/SS or the INS/SS to carry out integrated orbit and attitude determination; 1b) when the ground control point is less than M₂If the number of the matched feature points is more, namely the number of the matched feature points on two continuous images of the camera with the same view angle is more than M₃Firstly, adopting GNSS/INS/SS or INS/SS auxiliary beam method for block adjustment; 1c) when the ground control point is less than M₂If the number of the matched feature points is less, the number of the matched feature points on two continuous images of the camera with the same view angle is less than M₄Adopting GNSS/INS/SS or INS/SS to assist the adjustment of the area network with the beam method with relative attitude parameters; 1d) when the ground control point is less than M₂If the matched feature points are rare, the total number of the matched feature points on the 5 images of the same camera station is less than M₅And secondly, combining INS-assisted continuous sequential images based on GNSS/INS/SS or INS/SS to position and fix the attitude, namely using INS difference values in a short time as true values of relative attitude of the images, matching after N images are fixedly connected with N images, and then performing GNSS/INS/SS or INS/SS assisted beam method block network adjustment.

4. The method for autonomously determining the orbit and the attitude of the low orbit satellite based on the non-navigational satellite signals as claimed in claim 1, wherein: in step S2, for the 2) emergency period, there is a control point region,solving the high-precision space position and attitude of the satellite according to the following four conditions: 2a) when there are more ground control points, i.e. the number of control points on a single image is greater than M₁Secondly, a single-chip backward intersection auxiliary INS/SS is adopted to carry out integrated orbit and attitude determination; 2b) when the ground control point is less than M₂If the number of the matched feature points is more, namely the number of the matched feature points on two continuous images of the camera with the same view angle is more than M₃Firstly, adopting INS/SS auxiliary beam method block adjustment; 2c) when the ground control point is less than M₂If the number of the matched feature points is less, the number of the matched feature points on two continuous images of the camera with the same view angle is less than M₄Adopting INS/SS to assist adjustment of the area network with the beam method of relative attitude parameters; 2d) when the ground control point is less than M₂If the matched feature points are rare, the total number of the matched feature points on the 5 images of the same camera station is less than M₅And secondly, adopting INS-assisted continuous sequential images based on INS/SS to jointly position and fix the attitude, namely using the INS difference value of a short time as a true value of the relative attitude of the images, matching after N images are fixedly connected with N images, and then performing INS/SS assisted beam method block adjustment.

5. The method for autonomously determining the orbit and the attitude of the low orbit satellite based on the non-navigational satellite signals as claimed in claim 1, wherein: in step S2, for 3) a normal period, a control point-free region is solved for the high-precision spatial position and attitude of the satellite in two cases: 3a) when the SLAM is available, adopting GNSS/INS/SS/SLAM or INS/SS/SLAM to carry out integrated orbit and attitude determination and establishing a global geographic feature point library and continuously updating the global geographic feature point library; 3b) when the matched feature points are rare, namely the total number of the matched feature points on 5 images of the same camera station is less than M₆When SLAM is unavailable, INS based on GNSS/INS/SS or INS/SS is adopted to assist continuous sequential images to jointly position and fix the attitude, and the relative position and relative attitude between the images are considered to be known in a short period of time to form 'time continuous space separation images'.

6. The non-navigational satellite signal based low earth orbit satellite of claim 1The autonomous star orbit and attitude determination method is characterized by comprising the following steps: in step S2, for 4), in the emergency period, there is no control point region, the high-precision spatial position and attitude of the satellite are solved in two cases: 4a) when the SLAM is available, adopting INS/SS/SLAM to carry out integrated orbit and attitude determination and establishing a global geographic feature point library and continuously updating the global geographic feature point library; 4b) when the matched feature points are rare, namely the total number of the matched feature points on 5 images of the same camera station is less than M₆And when the SLAM is unavailable, the INS based on the INS/SS is adopted to assist the continuous sequential images to jointly position and fix the posture.

7. The method for autonomously determining the orbit and the attitude of the low orbit satellite based on the non-navigational satellite signals as claimed in any one of claims 3 to 6, wherein: the specific implementation mode of adopting the INS/SS to carry out integrated orbit and attitude determination is as follows,

the state variables are as follows:

wherein (phi)_Eφ_Nφ_U) Representing the northeast ENU attitude error (delta V) of the satellite-borne inertial sensor INS in the three-axis direction_EδV_NδV_U) Representing the speed error, (delta L delta lambda delta h) representing the position error of the satellite-borne inertial sensor INS, (epsilon)_Eε_Nε_U) Representing the gyro constant drift of the inertial device,

Representing a transformation matrix from the carrier coordinate system to the navigation coordinate system, the matrix F being a coefficient matrix, the matrix

W in W_εx，w_εy，w_εzAnd

zero mean random white noise for the gyroscope and accelerometer, respectively;

converting the three-axis error attitude angle into a mathematical platform coordinate system to obtain a misalignment angle attitude matrix, and taking the attitude information difference value measured by the attitude output and star sensor of the inertial navigation system as the observed quantity of a measurement equation, so that an observation equation Z is obtained₁＝H₁X+V₁The concrete expression is as follows:

8. The method for autonomously determining the orbit and the attitude of the low orbit satellite based on the non-navigational satellite signals as claimed in any one of claims 3 to 6, wherein: m₁Has a value of 4, M₂Is taken as 3, M₃Is 12 to 15, M₄Is 12 to 15, M₅Is 5 to 10, M₆The value of (1) is 5-10.